AI models comparison
30 major models, from the US to China through Europe. Powerful filters, adaptive view by your level, side-by-side comparison.
Up-to-date data·Verified 3 weeks ago(July 8, 2026)Next refresh : September 1, 2026View change history
🏆 The AI Podium
Official ranking · July 2026
Generalists
Best all-rounders
1
Kimi K2.6+2
Moonshot AI·88/100·$0.95/$4mit
Kimi K2.6 stands out thanks to strong reasoning (GPQA 91). Overall score: 88.1/100.
2
Claude Opus 4.8
Anthropic·85/100·$5/$25
Claude Opus 4.8 stands out thanks to a very large context window. Overall score: 84.71/100.
3
Gemini 3.1 Pro
Google·85/100·$2/$12
Gemini 3.1 Pro stands out thanks to a very large context window. Overall score: 84.5/100.
4
GPT-5.5-1
OpenAI·83/100·$5/$30
GPT-5.5 stands out thanks to a very large context window. Overall score: 83.3/100.
5
DeepSeek V3.2
DeepSeek·83/100·$0.28/$0.42mit
DeepSeek V3.2 stands out thanks to an exceptional price/performance ratio. Overall score: 82.65/100.
Code Specialists
SWE-bench + HumanEval
1
Claude Opus 4.8
Anthropic·87/100·$5/$25
Claude Opus 4.8 stands out thanks to a HumanEval score of 97%, a very recent release. Overall score: 87.25/100.
2
GPT-5.5
OpenAI·87/100·$5/$30
GPT-5.5 stands out thanks to a HumanEval score of 96%, a very recent release. Overall score: 87.15/100.
3
GPT-5
OpenAI·85/100·$1.25/$10
GPT-5 stands out thanks to a HumanEval score of 95%, a top SWE-bench score (89%). Overall score: 85.35/100.
4
Kimi K2.6
Moonshot AI·85/100·$0.95/$4mit
Kimi K2.6 stands out thanks to a HumanEval score of 92%, a very recent release. Overall score: 85.2/100.
5
Claude Opus 4.7-2
Anthropic·83/100·$5/$25
Claude Opus 4.7 stands out thanks to a HumanEval score of 97%, a top SWE-bench score (88%). Overall score: 83.25/100.
Local · Laptop
Light · < 16 GB RAM
1
Qwen3 0.6B
Alibaba·91/100apache-2.0
Qwen3 0.6B stands out thanks to a truly free Apache/MIT license. Overall score: 90.9/100.
2
Qwen3 8B
Alibaba·88/100apache-2.0
Qwen3 8B stands out thanks to a truly free Apache/MIT license. Overall score: 88.1/100.
3
Gemma 4 E4B
Google·88/100apache-2.0
Gemma 4 E4B stands out thanks to a truly free Apache/MIT license. Overall score: 87.7/100.
4
Gemma 4 12B
Google·86/100apache-2.0
Gemma 4 12B stands out thanks to a truly free Apache/MIT license. Overall score: 86.4/100.
5
Qwen3 4B
Alibaba·86/100apache-2.0
Qwen3 4B stands out thanks to a truly free Apache/MIT license. Overall score: 86/100.
Scroll horizontally
| Model | Release | Origin | Specialty | Context | Price | Size | Arena Elo | MMLU | HumanEval | SWE-Bench | Speed | Tags | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Doubao Seed 2.1 Pro ByteDance | Jun 2026 | 🇨🇳 China | 🧠 Generalist | 256K | $0.83 $/M | — | — | — | — | — | — | ||
Arbor Stability AI | Jun 2026 | 🇺🇸 United States | 3d-generation | 4K | 0€/m | — | — | — | — | — | — | OSSFree | |
GLM-5.2 Zhipu AI | Jun 2026 | 🇨🇳 China | 💻 Code | 1M | — | 754B Q4: 475 GB | — | — | — | — | — | OSS🏢 Server🤖 Agent | |
Kimi K2.7 Code Moonshot AI | Jun 2026 | 🇨🇳 China | 💻 Code | 256K | — | 1000B Q4: 629 GB | — | — | — | — | — | OSS🏢 Server🤖 Agent | |
| Jun 2026 | 🇺🇸 United States | 🔬 Reasoning | 1M | — | 550B Q4: 347 GB | — | — | — | — | — | OSS🏢 Server🤖 Agent | ||
Gemma 4 12B Google | Jun 2026 | 🇺🇸 United States | generalist | 256K | — | 12B Q4: 9 GB | — | — | — | — | — | OSS💻 Laptop🤖 Agent | |
Qwen3.7 Plus Alibaba | May 2026 | 🇨🇳 China | generalist | 1M | $0.40 $/M | — | — | — | — | — | — | ||
Claude Opus 4.8 Anthropic | May 2026 | 🇺🇸 United States | 🔬 Reasoning | 1M | $5.00 $/M | — | 1490 | 91.5% | 96.5% | 88.6% | 75 tok/s | FreeRGPD | |
Qwen3.7 Max Alibaba | May 2026 | 🇨🇳 China | generalist | 1M | $2.50 $/M | — | — | — | — | — | — | ||
Krea 2 Krea AI | May 2026 | 🇺🇸 United States | 🎨 Image generation | 4K | 8€/m | — | — | — | — | — | — | OSSFree | |
| Apr 2026 | 🇺🇸 United States | edge | 512K | — | 3B Q4: 4 GB | — | — | — | — | — | OSS💻 Laptop | ||
| Apr 2026 | 🇺🇸 United States | generalist | 512K | — | 8B Q4: 7 GB | — | — | — | — | — | OSS💻 Laptop🤖 Agent | ||
DeepSeek V4-Flash DeepSeek | Apr 2026 | 🇨🇳 China | generalist | 1M | — | 284B Q4: 180 GB | — | — | — | — | — | OSS🏢 Server🤖 Agent | |
GPT-5.5 OpenAI | Apr 2026 | 🇺🇸 United States | 🧠 Generalist | 1M | $5.00 $/M | — | — | — | — | 82.6% | — | Free | |
Qwen3.6 27B Alibaba | Apr 2026 | 🇨🇳 China | 💻 Code | 262K | — | 27B Q4: 19 GB | — | — | — | — | — | OSS🖥️ Desktop🤖 Agent | |
Kimi K2.6 Moonshot AI | Apr 2026 | 🇨🇳 China | 💻 Code | 262K | $0.95 $/M | — | — | — | — | — | — | OSSFree | |
Qwen3.6 35B-A3B Alibaba | Apr 2026 | 🇨🇳 China | 💻 Code | 262K | — | 35B Q4: 24 GB | — | — | — | — | — | OSS🖥️ Desktop🤖 Agent | |
Gemma 4 26B-A4B Google | Mar 2026 | 🇺🇸 United States | generalist | 256K | — | 26B Q4: 18 GB | — | — | — | — | — | OSS🖥️ Desktop🤖 Agent | |
Gemma 4 31B Google | Mar 2026 | 🇺🇸 United States | generalist | 256K | — | 31B Q4: 21 GB | — | — | — | — | — | OSS🖥️ Desktop🤖 Agent | |
| Mar 2026 | 🇺🇸 United States | generalist | 1M | — | 120B Q4: 77 GB | — | — | — | — | — | OSS🏢 Server🤖 Agent | ||
Gemma 4 E2B Google | Mar 2026 | 🇺🇸 United States | edge | 128K | — | 2B Q4: 3 GB | — | — | — | — | — | OSS💻 Laptop | |
Gemma 4 E4B Google | Mar 2026 | 🇺🇸 United States | edge | 128K | — | 4B Q4: 5 GB | — | — | — | — | — | OSS💻 Laptop | |
Gemini 3.1 Pro Google | Feb 2026 | 🇺🇸 United States | 🧠 Generalist | 1M | $2.00 $/M | — | — | — | — | 80.6% | 144 tok/s | Free | |
Qwen3.5 397B-A17B Alibaba | Feb 2026 | 🇨🇳 China | generalist | 262K | — | 397B Q4: 251 GB | — | — | — | — | — | OSS🏢 Server🤖 Agent | |
Seedance 2.0 ByteDance | Feb 2026 | 🇨🇳 China | 🎬 Video generation | 4K | 17€/m | — | — | — | — | — | — | Free | |
TRELLIS.2 Microsoft | Dec 2025 | 🇺🇸 United States | 3d-generation | 4K | 0€/m | — | — | — | — | — | — | OSSFree | |
Claude Opus 4.7 Anthropic | Dec 2025 | 🇺🇸 United States | 🔬 Reasoning | 500K | $5.00 $/M | — | 1390 | 89.5% | 94.2% | 72.4% | 75 tok/s | FreeRGPD | |
Nemotron 3 Nano 30B-A3B NVIDIA | Dec 2025 | 🇺🇸 United States | generalist | 1M | — | 30B Q4: 21 GB | — | — | — | — | — | OSS🖥️ Desktop🤖 Agent | |
Mistral Large 3 Mistral AI | Dec 2025 | 🇫🇷 France | 🧠 Generalist | 256K | $2.00 $/M | 675B Q4: 425 GB | 1335 | 84.5% | 88.0% | 50.0% | 90 tok/s | OSSFreeRGPD🏢 Server🤖 Agent | |
Ministral 3 14B Mistral AI | Dec 2025 | 🇫🇷 France | generalist | 256K | — | 14B Q4: 11 GB | — | — | — | — | — | OSS🖥️ Desktop🤖 Agent | |
Ministral 3 3B Mistral AI | Dec 2025 | 🇫🇷 France | edge | 256K | — | 3B Q4: 4 GB | — | — | — | — | — | OSS💻 Laptop | |
Ministral 3 8B Mistral AI | Dec 2025 | 🇫🇷 France | edge | 256K | — | 8B Q4: 7 GB | — | — | — | — | — | OSS💻 Laptop | |
DeepSeek V3.2 DeepSeek | Nov 2025 | 🇨🇳 China | 🧠 Generalist | 131K | $0.28 $/M | — | — | — | — | 67.8% | — | OSSFree | |
OLMo 3 7B Allen AI | Nov 2025 | 🇺🇸 United States | generalist | 66K | — | 7B Q4: 6 GB | — | — | — | — | — | OSS💻 Laptop | |
Gemini 3 Pro Google | Nov 2025 | 🇺🇸 United States | 🎯 Multimodal | 2M | $1.25 $/M | — | 1380 | 89.0% | 90.5% | 64.0% | 100 tok/s | FreeRGPD | |
Gemini 3 Flash Google | Nov 2025 | 🇺🇸 United States | 🧠 Generalist | 1M | $0.50 $/M | — | 1295 | 79.5% | 82.0% | — | 250 tok/s | FreeRGPD | |
Claude Sonnet 4.6 Anthropic | Oct 2025 | 🇺🇸 United States | 🧠 Generalist | 200K | $3.00 $/M | — | 1352 | 86.2% | 91.5% | 65.0% | 110 tok/s | FreeRGPD | |
Sora 2 OpenAI | Sep 2025 | 🇺🇸 United States | 🎬 Video generation | 4K | 20€/m | — | — | — | — | — | — | RGPD | |
Qwen 3 Max Alibaba | Sep 2025 | 🇨🇳 China | 🧠 Generalist | 256K | $0.78 $/M | — | 1325 | 86.0% | 87.5% | 38.0% | 80 tok/s | Free | |
GLM-5 Zhipu AI | Sep 2025 | 🇨🇳 China | 🧠 Generalist | 128K | $0.50 $/M | — | 1310 | 84.5% | 86.0% | 40.0% | 70 tok/s | OSSFree | |
Apertus 8B Swiss AI | Aug 2025 | 🌍 Switzerland | multilingual | 64K | — | 8B Q4: 7 GB | — | — | — | — | — | OSS💻 Laptop | |
GPT-5 OpenAI | Aug 2025 | 🇺🇸 United States | 🧠 Generalist | 400K | $1.25 $/M | — | 1395 | 90.8% | 92.0% | 68.5% | 90 tok/s | FreeRGPD | |
GPT-5 mini OpenAI | Aug 2025 | 🇺🇸 United States | 🧠 Generalist | 200K | $0.15 $/M | — | 1290 | 82.5% | 85.0% | — | 200 tok/s | FreeRGPD | |
gpt-oss-120b OpenAI | Aug 2025 | 🇺🇸 United States | generalist | 128K | — | 117B Q4: 75 GB | — | — | — | — | — | OSS🏢 Server🤖 Agent | |
gpt-oss-20b OpenAI | Aug 2025 | 🇺🇸 United States | generalist | 128K | — | 21B Q4: 15 GB | — | — | — | — | — | OSS🖥️ Desktop🤖 Agent | |
Qwen3-Coder 480B Alibaba | Jul 2025 | 🇨🇳 China | 💻 Code | 256K | — | 480B Q4: 303 GB | — | — | — | — | — | OSS🏢 Server🤖 Agent | |
GLM-4.5 Z.ai | Jul 2025 | 🇨🇳 China | generalist | 128K | — | 355B Q4: 225 GB | — | — | — | — | — | OSS🏢 Server🤖 Agent | |
GLM-4.5 Air Z.ai | Jul 2025 | 🇨🇳 China | generalist | 128K | — | 106B Q4: 68 GB | — | — | — | — | — | OSS🏢 Server🤖 Agent | |
Kimi K2 Moonshot AI | Jul 2025 | 🇨🇳 China | generalist | 128K | — | 1026B Q4: 645 GB | — | — | — | — | — | OSS🏢 Server🤖 Agent | |
Grok 4 xAI | Jul 2025 | 🇺🇸 United States | 🧠 Generalist | 256K | $5.00 $/M | — | 1360 | 87.0% | 88.5% | 60.0% | 60 tok/s | ||
SmolLM3 3B Hugging Face | Jul 2025 | 🇺🇸 United States | edge | 64K | — | 3.1B Q4: 4 GB | — | — | — | — | — | OSS💻 Laptop | |
Mistral Small 3.2 Mistral AI | Jun 2025 | 🇫🇷 France | generalist | 128K | — | 24B Q4: 17 GB | — | — | — | — | — | OSS🖥️ Desktop🤖 Agent | |
ElevenLabs v3 ElevenLabs | Jun 2025 | 🇺🇸 United States | 🎙️ Voice synthesis | 10K | 5€/m | — | — | — | — | — | — | FreeRGPD | |
Magistral Small Mistral AI | Jun 2025 | 🇫🇷 France | 🔬 Reasoning | 40K | — | 24B Q4: 17 GB | — | — | — | — | — | OSS🖥️ Desktop | |
Gemma 3n E4B Google | Jun 2025 | 🇺🇸 United States | edge | 32K | — | 8B Q4: 7 GB | — | — | — | — | — | OSS💻 Laptop | |
DeepSeek R2 DeepSeek | May 2025 | 🇨🇳 China | 🔬 Reasoning | 128K | $0.55 $/M | — | 1370 | 89.0% | 92.5% | 49.2% | 25 tok/s | OSSFree | |
Devstral Small Mistral AI | May 2025 | 🇫🇷 France | 💻 Code | 128K | — | 24B Q4: 17 GB | — | — | — | — | — | OSS🖥️ Desktop🤖 Agent | |
Qwen3 0.6B Alibaba | Apr 2025 | 🇨🇳 China | edge | 32K | — | 600M Q4: 2 GB | — | — | — | — | — | OSS💻 Laptop | |
Qwen3 14B Alibaba | Apr 2025 | 🇨🇳 China | generalist | 128K | — | 14.8B Q4: 11 GB | — | — | — | — | — | OSS🖥️ Desktop🤖 Agent | |
Qwen3 235B-A22B Alibaba | Apr 2025 | 🇨🇳 China | generalist | 128K | — | 235B Q4: 149 GB | — | — | — | — | — | OSS🏢 Server🤖 Agent | |
Qwen3 32B Alibaba | Apr 2025 | 🇨🇳 China | generalist | 128K | — | 32.8B Q4: 23 GB | — | — | — | — | — | OSS🖥️ Desktop🤖 Agent | |
Qwen3 4B Alibaba | Apr 2025 | 🇨🇳 China | edge | 32K | — | 4.02B Q4: 5 GB | — | — | — | — | — | OSS💻 Laptop | |
Qwen3 8B Alibaba | Apr 2025 | 🇨🇳 China | generalist | 128K | — | 8.2B Q4: 7 GB | — | — | — | — | — | OSS💻 Laptop | |
o3 OpenAI | Apr 2025 | 🇺🇸 United States | 🔬 Reasoning | 200K | $2.00 $/M | — | 1410 | 88.5% | 96.0% | 71.7% | 30 tok/s | RGPD | |
| Apr 2025 | 🇺🇸 United States | enterprise | 128K | — | 8.1B Q4: 7 GB | — | — | — | — | — | OSS💻 Laptop🤖 Agent | ||
Phi-4 Reasoning Microsoft | Apr 2025 | 🇺🇸 United States | 🔬 Reasoning | 32K | — | 14.7B Q4: 11 GB | — | — | — | — | — | OSS🖥️ Desktop | |
Llama 4 Behemoth Meta | Apr 2025 | 🇺🇸 United States | 🧠 Generalist | 1M | $3.00 $/M | — | 1330 | 87.0% | 88.5% | 45.0% | 40 tok/s | OSSFree | |
Llama 4 Scout Meta | Apr 2025 | 🇺🇸 United States | 🧠 Generalist | 10M | $0.40 $/M | 109B Q4: 70 GB | 1290 | 80.0% | 84.0% | — | 90 tok/s | OSSFree🏢 Server | |
Midjourney v7 Midjourney | Apr 2025 | 🇺🇸 United States | 🎨 Image generation | 4K | 10€/m | — | — | — | — | — | — | RGPD | |
Llama 4 Maverick Meta | Mar 2025 | 🇺🇸 United States | generalist | 1M | — | 400B Q4: 253 GB | — | — | — | — | — | OSS🏢 Server | |
Runway Gen-4 Runway | Mar 2025 | 🇺🇸 United States | 🎬 Video generation | 4K | 12€/m | — | — | — | — | — | — | FreeRGPD | |
OLMo 2 32B Allen AI | Mar 2025 | 🇺🇸 United States | research | 4K | — | 32B Q4: 22 GB | — | — | — | — | — | OSS🖥️ Desktop | |
Gemma 3 1B Google | Mar 2025 | 🇺🇸 United States | edge | 32K | — | 1B Q4: 3 GB | — | — | — | — | — | OSS💻 Laptop | |
QwQ 32B Alibaba | Mar 2025 | 🇨🇳 China | 🔬 Reasoning | 128K | — | 32.8B Q4: 23 GB | — | — | — | — | — | OSS🖥️ Desktop | |
Gemma 3 12B Google | Feb 2025 | 🇺🇸 United States | generalist | 128K | — | 12B Q4: 9 GB | — | — | — | — | — | OSS💻 Laptop | |
Gemma 3 27B Google | Feb 2025 | 🇺🇸 United States | generalist | 128K | — | 27B Q4: 19 GB | — | — | — | — | — | OSS🖥️ Desktop | |
DeepSeek Coder V3 DeepSeek | Feb 2025 | 🇨🇳 China | 💻 Code | 128K | $0.14 $/M | — | 1280 | — | 90.2% | 49.0% | 80 tok/s | OSSFree | |
Gemma 3 4B Google | Feb 2025 | 🇺🇸 United States | edge | 128K | — | 4B Q4: 5 GB | — | — | — | — | — | OSS💻 Laptop | |
Phi-4 Mini Microsoft | Feb 2025 | 🇺🇸 United States | edge | 128K | — | 3.8B Q4: 4 GB | — | — | — | — | — | OSS💻 Laptop | |
Lucie 7B OpenLLM France | Feb 2025 | 🇫🇷 France | multilingual | 32K | — | 6.7B Q4: 6 GB | — | — | — | — | — | OSS💻 Laptop | |
Mistral Small 3 Mistral AI | Jan 2025 | 🇫🇷 France | 🧠 Generalist | 32K | $0.20 $/M | — | 1235 | 75.0% | 80.0% | — | 150 tok/s | OSSFreeRGPD | |
DeepSeek-R1 DeepSeek | Jan 2025 | 🇨🇳 China | 🔬 Reasoning | 128K | — | 671B Q4: 423 GB | — | — | — | — | — | OSS🏢 Server | |
DeepSeek-R1 Distill 32B DeepSeek | Jan 2025 | 🇨🇳 China | 🔬 Reasoning | 128K | — | 32.8B Q4: 23 GB | — | — | — | — | — | OSS🖥️ Desktop | |
DeepSeek-R1 Distill 8B DeepSeek | Jan 2025 | 🇨🇳 China | 🔬 Reasoning | 128K | — | 8.03B Q4: 7 GB | — | — | — | — | — | OSS💻 Laptop | |
Codestral Mistral AI | Jan 2025 | 🇫🇷 France | 💻 Code | 256K | $0.30 $/M | 22.2B Q4: 16 GB | 1240 | — | 81.1% | 35.0% | 130 tok/s | OSSFreeRGPD🖥️ Desktop | |
InternLM3 8B Shanghai AI Lab | Jan 2025 | 🇨🇳 China | generalist | 32K | — | 8.8B Q4: 7 GB | — | — | — | — | — | OSS💻 Laptop | |
DeepSeek V3 DeepSeek | Dec 2024 | 🇨🇳 China | 🧠 Generalist | 128K | $0.27 $/M | 685B Q4: 431 GB | 1340 | 87.5% | 89.0% | 42.0% | 60 tok/s | OSSFree🏢 Server | |
Falcon 3 10B TII | Dec 2024 | 🇦🇪 United Arab Emirates | generalist | 32K | — | 10.3B Q4: 8 GB | — | — | — | — | — | OSS💻 Laptop | |
Command R7B Cohere | Dec 2024 | 🇨🇦 Canada | rag | 128K | — | 8B Q4: 7 GB | — | — | — | — | — | OSS💻 Laptop🤖 Agent | |
Phi-4 Microsoft | Dec 2024 | 🇺🇸 United States | generalist | 16K | — | 14.7B Q4: 11 GB | — | — | — | — | — | OSS🖥️ Desktop | |
Llama 3.3 70B Meta | Nov 2024 | 🇺🇸 United States | generalist | 128K | — | 70.6B Q4: 46 GB | — | — | — | — | — | OSS🗄️ Workstation | |
EuroLLM 9B UTTER Project | Nov 2024 | 🇪🇺 European Union | multilingual | 4K | — | 9.15B Q4: 8 GB | — | — | — | — | — | OSS💻 Laptop | |
Qwen2.5-Coder 32B Alibaba | Nov 2024 | 🇨🇳 China | 💻 Code | 128K | — | 32.8B Q4: 23 GB | — | — | — | — | — | OSS🖥️ Desktop🤖 Agent | |
Yi-Lightning 01.AI | Oct 2024 | 🇨🇳 China | 🧠 Generalist | 32K | $0.14 $/M | — | 1287 | 81.5% | 82.0% | — | 180 tok/s | Free | |
Aya Expanse 32B Cohere | Oct 2024 | 🇨🇦 Canada | multilingual | 128K | — | 32.3B Q4: 22 GB | — | — | — | — | — | OSS🖥️ Desktop | |
Stable Diffusion 3.5 Stability AI | Oct 2024 | 🇬🇧 United Kingdom | 🎨 Image generation | 4K | — | — | — | — | — | — | — | OSSFreeRGPD | |
Ministral 8B Mistral AI | Oct 2024 | 🇫🇷 France | edge | 128K | — | 8.02B Q4: 7 GB | — | — | — | — | — | OSS💻 Laptop | |
Llama 3.1 Nemotron 70B NVIDIA | Oct 2024 | 🇺🇸 United States | generalist | 128K | — | 70.6B Q4: 46 GB | — | — | — | — | — | OSS🗄️ Workstation | |
OpenAI Voice (Realtime) OpenAI | Sep 2024 | 🇺🇸 United States | 🎙️ Voice synthesis | 128K | $100.00 $/M | — | — | — | — | — | — | RGPD | |
Llama 3.2 1B Meta | Sep 2024 | 🇺🇸 United States | edge | 128K | — | 1.24B Q4: 3 GB | — | — | — | — | — | OSS💻 Laptop | |
Llama 3.2 3B Meta | Sep 2024 | 🇺🇸 United States | edge | 128K | — | 3.21B Q4: 4 GB | — | — | — | — | — | OSS💻 Laptop | |
Qwen2.5 72B Alibaba | Sep 2024 | 🇨🇳 China | generalist | 128K | — | 72.7B Q4: 48 GB | — | — | — | — | — | OSS🗄️ Workstation | |
Command R+ Cohere | Aug 2024 | 🇨🇦 Canada | rag | 128K | — | 104B Q4: 67 GB | — | — | — | — | — | OSS🏢 Server🤖 Agent | |
FLUX.1 Pro Black Forest Labs | Jul 2024 | 🇩🇪 Germany | 🎨 Image generation | 4K | $50.00 $/M | — | — | — | — | — | — | FreeRGPD | |
Hermes 3 8B Nous Research | Jul 2024 | 🇺🇸 United States | generalist | 128K | — | 8.03B Q4: 7 GB | — | — | — | — | — | OSS💻 Laptop🤖 Agent | |
Llama 3.1 8B Meta | Jul 2024 | 🇺🇸 United States | generalist | 128K | — | 8.03B Q4: 7 GB | — | — | — | — | — | OSS💻 Laptop | |
Mistral Nemo 12B Mistral AI | Jul 2024 | 🇫🇷 France | generalist | 128K | — | 12.2B Q4: 10 GB | — | — | — | — | — | OSS💻 Laptop | |
DeepSeek-Coder-V2 Lite DeepSeek | Jun 2024 | 🇨🇳 China | 💻 Code | 128K | — | 15.7B Q4: 12 GB | — | — | — | — | — | OSS🖥️ Desktop | |
Mistral 7B Mistral AI | May 2024 | 🇫🇷 France | generalist | 32K | — | 7.25B Q4: 7 GB | — | — | — | — | — | OSS💻 Laptop | |
Mixtral 8x22B Mistral AI | Apr 2024 | 🇫🇷 France | 🧠 Generalist | 64K | $0.60 $/M | — | 1145 | 77.5% | 76.0% | — | 100 tok/s | OSSFreeRGPD | |
CodeGemma 7B Google | Mar 2024 | 🇺🇸 United States | 💻 Code | 8K | — | 8.5B Q4: 7 GB | — | — | — | — | — | OSS💻 Laptop | |
StarCoder 2 BigCode | Feb 2024 | 🇪🇺 Europe (collective) | 💻 Code | 16K | — | 16B Q4: 12 GB | — | — | 46.3% | — | 200 tok/s | OSSFreeRGPD🖥️ Desktop | |
Mixtral 8x7B Mistral AI | Dec 2023 | 🇫🇷 France | generalist | 32K | — | 46.7B Q4: 31 GB | — | — | — | — | — | OSS🗄️ Workstation | |
DALL-E 3 OpenAI | Oct 2023 | 🇺🇸 United States | 🎨 Image generation | 4K | $40.00 $/M | — | — | — | — | — | — | RGPD | |
Falcon 180B TII | Sep 2023 | 🇦🇪 UAE | 🧠 Generalist | 4K | — | — | — | 70.5% | 35.0% | — | 30 tok/s | OSSFree |
Indicators legend
🇫🇷
French quality: rating out of 5 based on fluency, grammar, and cultural nuances.
🧠
Reasoning depth: ability to chain complex logical steps.
⚡
Perceived speed: feeling of responsiveness in use, combines latency and throughput.
🎨
Creativity: quality of imaginative generation, originality of responses.
🔬
Safety filters level: 1 = very permissive, 5 = very strict (frequent refusals).
★★★
Ratings 1 to 5 stars, ★ = poor, ★★★★★ = excellent. Indicative qualitative assessments.
Frequently asked questions
What is the best AI model in 2026?▾
It depends on your use case. For reasoning and French writing quality, Claude Opus 4.7 leads. For multimodal and very long context, Gemini 3 Pro. For value, Claude Sonnet 4.6 or Mistral Large 3 (European, GDPR). For open-source, DeepSeek V3 or Llama 4.
Which AI is GDPR-compliant for a European company?▾
Mistral Large 3 (France-hosted), Mistral Small 3 (open-source, self-hostable), Codestral, and European versions of US models (Claude via Anthropic EU, Gemini via Google Cloud EU). Avoid Grok 4 and most Chinese models for sensitive European data.
Which free AI to use?▾
24 of the 30 models here have a free tier. For daily use: Claude Sonnet 4.6 (claude.ai), GPT-5 (chatgpt.com), Gemini 3 Pro (gemini.google.com). For free self-hosting: DeepSeek V3, Llama 4, Mistral Small 3. Filter by "Free tier" above to see all free models.
What is Arena Elo and how to interpret it?▾
Arena Elo (LMSYS Chatbot Arena) is a score from blind human comparisons between models. Users vote without knowing which model answered. It's one of the most reliable benchmarks because it reflects real human preference. Reference: 1200 = good, 1300 = very good, 1400+ = world-class.
Open-source vs proprietary: which to choose?▾
Open-source if you want: self-hosting, fine-tuning, data control, no vendor lock-in. Proprietary if you want: maximum performance, no infra management, enterprise support. Many organizations use both: proprietary for the frontier, open-source for volume and sensitive data.